Litcius/Paper detail

A Feature Cum Intensity Based SSIM Optimised Hybrid Image Registration Technique

Taruna Kumari, Vikrant Guleria, Poonam Syal, Ashwani Kumar Aggarwal

202115 citationsDOI

Abstract

The hybrid image registration method is proposed to align two images based on the features or corresponding intensity information present in the images. The motivation behind the proposed work is that there is no one method or algorithm available that is suitable for any kind of images. SURF feature-based algorithm is used to extract, match, and describe the features present in the image. <tex>$1+1$</tex> evolutionary and regular step gradient descent algorithm is used for intensity information present in the images. The performance parameter to evaluate registration accuracy used in the proposed work are SSIM, MSE, PSNR, IQI, PCC and SSD which shows improvement in prior and post registration results.

Topics & Concepts

Artificial intelligenceImage registrationFeature (linguistics)Computer scienceComputer visionImage (mathematics)Gradient descentPattern recognition (psychology)Intensity (physics)Feature extractionArtificial neural networkPhysicsQuantum mechanicsPhilosophyLinguisticsMedical Image Segmentation TechniquesAdvanced Image and Video Retrieval TechniquesRobotics and Sensor-Based Localization
A Feature Cum Intensity Based SSIM Optimised Hybrid Image Registration Technique | Litcius